基于LSTM神經(jīng)網(wǎng)絡(luò)的電池SOH估算

??循環(huán)神經(jīng)網(wǎng)絡(luò)（Recurrent Neural Network，RNN）與BP神經(jīng)網(wǎng)絡(luò)不同，RNN網(wǎng)絡(luò)不僅考慮前一時刻的輸入，同時還賦予網(wǎng)絡(luò)對前面時刻信息的記憶能力。盡管RNN網(wǎng)絡(luò)具有較高的精度，但其存在著梯度消失的問題。對此，出現(xiàn)了一系列改進的RNN網(wǎng)絡(luò)，而LSTM神經(jīng)網(wǎng)絡(luò)就是其中改進效果最好的一種。基于LSTM神經(jīng)網(wǎng)絡(luò)的電池SOH估算方法具體如下所示：
??（1）鋰離子電池循環(huán)壽命數(shù)據(jù)
??數(shù)據(jù)來源于NASA研究中心搭建的鋰離子電池測試平臺，選取5號鋰離子電池（額定容量為2Ah）。循環(huán)測試實驗在室溫下進行：以1.5A的恒流電流為電池充電，直到充電截止電壓（4.2V），以恒壓電流充電，直到充電電流降至20mA；在2A恒流（CC）模式下放電，直至電池分別降至2.7V、2.5V、2.2V、2.5V。當(dāng)電池達(dá)到EOL（end-of-life, EOL）標(biāo)準(zhǔn)時，實驗停止，額定容量下降30%。
??（2）數(shù)據(jù)預(yù)處理
??利用最大最小值歸一化方法將數(shù)據(jù)放縮到0和1之間。

??式中，xmax是輸入數(shù)據(jù)x的最大值，xmin是輸入數(shù)據(jù)x的最小值。
??（3）仿真分析
??將NASA鋰離子電池實驗數(shù)據(jù)中的充電容量作為模型輸入，模型的輸出為以放電容量為參考的電池 SOH。經(jīng)過訓(xùn)練，神經(jīng)網(wǎng)絡(luò)可以迭代得到網(wǎng)絡(luò)權(quán)重及偏置具體參數(shù)。

??從圖中可以看出，模型對于電池容量衰減的整體趨勢預(yù)測是非常精準(zhǔn)的。其具體代碼如下所示：

function pre_data = LSTM_main(d,h,train_data,test_data) %% 預(yù)處理 lag = 8; % d = 51; [train_input,train_output] = LSTM_data_process(d,train_data,lag); %數(shù)據(jù)處理[train_input,min_input,max_input,train_output,min_output,max_output] = premnmx(train_input',train_output');input_length = size(train_input,1); %樣本輸入長度 output_length = size(train_output,1); %樣本輸出長度 train_num = size(train_input,2); %訓(xùn)練樣本個數(shù) test_num = size(test_data,2); %測試樣本個數(shù) %% 網(wǎng)絡(luò)參數(shù)初始化 % 結(jié)點數(shù)設(shè)置 input_num = input_length; cell_num = 10; output_num = output_length; % 網(wǎng)絡(luò)中門的偏置 bias_input_gate = rand(1,cell_num); bias_forget_gate = rand(1,cell_num); bias_output_gate = rand(1,cell_num); %網(wǎng)絡(luò)權(quán)重初始化 ab = 20; weight_input_x = rand(input_num,cell_num)/ab; weight_input_h = rand(output_num,cell_num)/ab; weight_inputgate_x = rand(input_num,cell_num)/ab; weight_inputgate_h = rand(cell_num,cell_num)/ab; weight_forgetgate_x = rand(input_num,cell_num)/ab; weight_forgetgate_h = rand(cell_num,cell_num)/ab; weight_outputgate_x = rand(input_num,cell_num)/ab; weight_outputgate_h = rand(cell_num,cell_num)/ab; %hidden_output權(quán)重 weight_preh_h = rand(cell_num,output_num); %網(wǎng)絡(luò)狀態(tài)初始化 cost_gate = 1e-6; h_state = rand(output_num,train_num+test_num); cell_state = rand(cell_num,train_num+test_num); %% 網(wǎng)絡(luò)訓(xùn)練學(xué)習(xí) for iter = 1:3000 %迭代次數(shù)yita = 0.01; %每次迭代權(quán)重調(diào)整比例for m = 1:train_num%前饋部分if(m==1)gate = tanh(train_input(:,m)' * weight_input_x);input_gate_input = train_input(:,m)' * weight_inputgate_x + bias_input_gate;output_gate_input = train_input(:,m)' * weight_outputgate_x + bias_output_gate;for n = 1:cell_numinput_gate(1,n) = 1 / (1 + exp(-input_gate_input(1,n)));%輸入門output_gate(1,n) = 1 / (1 + exp(-output_gate_input(1,n)));%輸出門%sigmoid函數(shù)endforget_gate = zeros(1,cell_num);forget_gate_input = zeros(1,cell_num);cell_state(:,m) = (input_gate .* gate)';elsegate = tanh(train_input(:,m)' * weight_input_x + h_state(:,m-1)' * weight_input_h);input_gate_input = train_input(:,m)' * weight_inputgate_x + cell_state(:,m-1)' * weight_inputgate_h + bias_input_gate;forget_gate_input = train_input(:,m)' * weight_forgetgate_x + cell_state(:,m-1)' * weight_forgetgate_h + bias_forget_gate;output_gate_input = train_input(:,m)' * weight_outputgate_x + cell_state(:,m-1)' * weight_outputgate_h + bias_output_gate;for n = 1:cell_numinput_gate(1,n) = 1/(1+exp(-input_gate_input(1,n)));forget_gate(1,n) = 1/(1+exp(-forget_gate_input(1,n)));output_gate(1,n) = 1/(1+exp(-output_gate_input(1,n)));endcell_state(:,m) = (input_gate .* gate + cell_state(:,m-1)' .* forget_gate)'; endpre_h_state = tanh(cell_state(:,m)') .* output_gate;h_state(:,m) = (pre_h_state * weight_preh_h)';%誤差計算Error = h_state(:,m) - train_output(:,m);Error_Cost(1,iter)=sum(Error.^2); %誤差的平方和（4個點的平方和）if(Error_Cost(1,iter)<cost_gate) %判斷是否滿足誤差最小條件flag = 1;break;else [ weight_input_x,...weight_input_h,...weight_inputgate_x,...weight_inputgate_h,...weight_forgetgate_x,...weight_forgetgate_h,...weight_outputgate_x,...weight_outputgate_h,...weight_preh_h ] = LSTM_updata_weight(m,yita,Error,...weight_input_x,...weight_input_h,...weight_inputgate_x,...weight_inputgate_h,...weight_forgetgate_x,...weight_forgetgate_h,...weight_outputgate_x,...weight_outputgate_h,...weight_preh_h,...cell_state,h_state,...input_gate,forget_gate,...output_gate,gate,...train_input,pre_h_state,...input_gate_input,...output_gate_input,...forget_gate_input,input_num,cell_num);endendif(Error_Cost(1,iter)<cost_gate)break;end end%% 測試階段 %數(shù)據(jù)加載 test_input = train_data(end-lag+1:end); test_input = tramnmx(test_input',min_input,max_input); % test_input = mapminmax('apply',test_input',ps_input);%前饋 for m = train_num + 1:train_num + test_numgate = tanh(test_input' * weight_input_x + h_state(:,m-1)' * weight_input_h);input_gate_input = test_input' * weight_inputgate_x + h_state(:,m-1)' * weight_inputgate_h + bias_input_gate;forget_gate_input = test_input' * weight_forgetgate_x + h_state(:,m-1)' * weight_forgetgate_h + bias_forget_gate;output_gate_input = test_input' * weight_outputgate_x + h_state(:,m-1)' * weight_outputgate_h + bias_output_gate;for n = 1:cell_numinput_gate(1,n) = 1/(1+exp(-input_gate_input(1,n)));forget_gate(1,n) = 1/(1+exp(-forget_gate_input(1,n)));output_gate(1,n) = 1/(1+exp(-output_gate_input(1,n))); endcell_state(:,m) = (input_gate .* gate + cell_state(:,m-1)' .* forget_gate)'; pre_h_state = tanh(cell_state(:,m)') .* output_gate;h_state(:,m) = (pre_h_state * weight_preh_h)';% 將當(dāng)前預(yù)測點作為下一步輸入數(shù)據(jù)test_input = postmnmx(test_input,min_input,max_input);now_prepoint = postmnmx(h_state(:,m),min_output,max_output);%test_input = mapminmax('reverse',test_input,ps_input);test_input = [test_input(2:end); now_prepoint];test_input = tramnmx(test_input,min_input,max_input); endpre_data = postmnmx(h_state(:,train_num + h:h:train_num + test_num),min_output,max_output);all_pre = postmnmx(h_state(:,1:h:train_num + test_num),min_output,max_output);% 畫圖 figure title('LSTM預(yù)測') hold on plot(1:size([train_data test_data],2),[train_data test_data], 'o-', 'color','r', 'linewidth', 1); plot(size(train_data,2) + h:h:size([train_data test_data],2),pre_data, '*-','color','b','linewidth', 1); plot([size(train_data,2) size(train_data,2)],[-0.01 0.01],'g-','LineWidth',4); legend({ '真實值', '預(yù)測值'}); end

??想了解更過相關(guān)仿真，可以關(guān)注我的微信公眾號。

總結(jié)

以上是生活随笔為你收集整理的电池SOH仿真系列-基于LSTM神经网络的电池SOH估算方法的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。